A Combined MPI-CUDA Parallel Solution of Linear and Nonlinear Poisson-Boltzmann Equation

Figure 9

Execution time per iteration of the linear stencil versus number of GPU cards. Done on Cluster 2 and resource 3.